GPT-5.6 Luna vs Phi-3-medium-instruct (128k)
Phi-3-medium-instruct (128k) is cheaper on output tokens, while GPT-5.6 Luna offers a larger context window. Choose GPT-5.6 Luna or Phi-3-medium-instruct (128k) based on the trade-off between cost, context, and the benchmarks that matter for your use case.
| Spec | GPT-5.6 Luna | Phi-3-medium-instruct (128k) |
|---|---|---|
| Provider | Azure Cognitive Services | Azure Cognitive Services |
| Input / 1M tokens | $1.00 | $0.17 |
| Output / 1M tokens | $6.00 | $0.68 |
| Context window | 1.1M | 128K |
| Parameters | — | — |
| Open weights | No | Yes |
| Released | Jul 2026 | Apr 2024 |
FAQ
Pricing is indicative — confirm with the provider before production use. Highlighted values indicate the better figure for that row.